
import pandas as pd
import seaborn as sns
import plotly.express as px
import plotly.graph_objects as go

# Comment this if the data visualisations doesn't work on your side
%matplotlib inline

dfMovies = pd.read_csv('moviedataset3_11_21.csv')
dfReviews = pd.read_csv('reviewDataSet.csv')


dfMovies.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 14213 entries, 0 to 14212
Data columns (total 13 columns):
 #   Column       Non-Null Count  Dtype 
---  ------       --------------  ----- 
 0   title        14213 non-null  object
 1   age_rating   14213 non-null  object
 2   rating       14213 non-null  object
 3   rank         14213 non-null  int64 
 4   genre        14213 non-null  object
 5   director     14213 non-null  object
 6   year         14213 non-null  int64 
 7   producer     13941 non-null  object
 8   actor        14213 non-null  object
 9   runtime      14213 non-null  int64 
 10  description  14210 non-null  object
 11  img          14213 non-null  object
 12  url          14213 non-null  object
dtypes: int64(3), object(10)
memory usage: 1.4+ MB


dfReviews.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 584150 entries, 0 to 584149
Data columns (total 5 columns):
 #   Column  Non-Null Count   Dtype 
---  ------  --------------   ----- 
 0   title   584150 non-null  object
 1   user    584148 non-null  object
 2   type    584150 non-null  object
 3   grade   584150 non-null  int64 
 4   review  584150 non-null  object
dtypes: int64(1), object(4)
memory usage: 22.3+ MB


fig = px.pie(dfMovies, names=dfMovies["genre"].value_counts().index, values=dfMovies["genre"].value_counts())
fig.update_traces(hoverinfo='value', textinfo='label+percent')
fig.update_layout(autosize=False,width=1000,height=1000,)
fig.show();


fig = px.bar(dfMovies, x=dfMovies["age_rating"].value_counts().index, y=dfMovies["age_rating"].value_counts())
fig.update_layout(autosize=False,width=1000,height=300,)
fig.show();


fig = px.histogram(dfMovies, x=dfMovies["year"].value_counts().index, y=dfMovies["year"].value_counts(),nbins=130)
fig.update_layout(autosize=False,width=1000,height=300,)
fig.show();


prodCounts=pd.DataFrame(dfMovies["producer"].value_counts()).head(30)
test_tree = go.Figure(go.Treemap(
    labels =  prodCounts["producer"].index,
    parents=[""]*len(prodCounts["producer"].index),
    values =  prodCounts["producer"].values,
    textinfo = "label+value"
))
test_tree.show();


fig = px.box(dfMovies, x="runtime", title="Boxplot representativo del tiempo de ejecución.")
fig.show();


actorCounts=pd.DataFrame(dfMovies["actor"].value_counts())[1:26]
fig = px.pie(actorCounts, names=actorCounts["actor"].index, values=actorCounts["actor"])
fig.update_layout(autosize=False,width=900,height=490,)
fig.update_traces(hoverinfo='value', textinfo='label+percent')

fig.show();


directorCounts=pd.DataFrame(dfMovies["director"].value_counts()).head(25)
z=list(range(1, 26))
fig = go.Figure(data=[go.Bar(
    x=directorCounts["director"].index,
    y=directorCounts["director"],
     marker=dict(color = z, colorscale='Portland'))
])
fig.update_layout(title_text='Top 25 actores');


fig = px.box(dfReviews, x="grade",y="type")
fig.show()


reviewcounts=pd.DataFrame(dfReviews["title"].value_counts()).head(40)
test_tree = go.Figure(go.Treemap(
    labels =  reviewcounts["title"].index,
    parents=[""]*len(reviewcounts["title"].index),
    values =  reviewcounts["title"].values,
    textinfo = "label+value"
))
test_tree.show();


print("Reviews duplicadas")
print(dfReviews.duplicated().value_counts())
print("")
print("Peliculas duplicadas")
print(dfMovies.duplicated().value_counts())

Reviews duplicadas
False    583727
dtype: int64

Peliculas duplicadas
False    14213
dtype: int64


dfReviews=dfReviews.drop_duplicates()
dfReviews.duplicated().value_counts()

False    583727
dtype: int64


print("Edades antes de procesado",dfMovies["age_rating"].unique() )
df_mod=dfMovies.replace(" | Approved", 15)
df_mod=df_mod.replace(" | R", 18)
df_mod=df_mod.replace(' | TV-G', 0)
df_mod=df_mod.replace(' | TV-PG', 15)
df_mod=df_mod.replace(' | PG', 15)
df_mod=df_mod.replace(' | G', 0)
df_mod=df_mod.replace(' | Passed', 0)
df_mod=df_mod.replace(' | Not Rated', -5)
df_mod=df_mod.replace(' | PG-13', 13)
df_mod=df_mod.replace(' | PG-13`', 13)
df_mod=df_mod.replace(' | GP', 0)
df_mod=df_mod.replace(' | M', 15)
df_mod=df_mod.replace(' | M/PG', 15)
df_mod=df_mod.replace(' | TV-MA', 17)
df_mod=df_mod.replace(' | Unrated', -5)
df_mod=df_mod.replace('Not Rated', -5)
df_mod=df_mod.replace(' | TV-14', 14)
df_mod=df_mod.replace(' | NC-17', 17)
df_mod=df_mod.replace(' | NR', -5)
df_mod=df_mod.replace(' | Open', 14)
df_mod=df_mod.replace(' | X', 18)
df_mod=df_mod.replace(' | MA-17', 14)
df_mod=df_mod.replace(' | PG--13', 13)
df_mod=df_mod.replace(' | TV-Y7-FV', 7)
df_mod=df_mod.replace(' | TV-Y7', 7)
print()
print()
print("Edades despues de procesado",df_mod["age_rating"].unique() )

Edades antes de procesado [' | Approved' ' | R' ' | TV-G' ' | TV-PG' ' | PG' ' | G' ' | Passed'
 ' | Not Rated' ' | PG-13' ' | GP' ' | M' ' | M/PG' ' | TV-MA'
 ' | Unrated' 'Not Rated' ' | TV-14' ' | NC-17' ' | NR' ' | Open' ' | X'
 ' | MA-17' ' | PG--13' ' | PG-13`' ' | TV-Y7-FV' ' | TV-Y7']


Edades despues de procesado [15 18  0 -5 13 17 14  7]


fig = px.bar(df_mod, x=df_mod["age_rating"].value_counts().index, y=df_mod["age_rating"].value_counts())
fig.update_layout(autosize=False,width=700,height=500,)
fig.show();


print(df_mod['age_rating'].loc[df_mod['age_rating'] > -1].value_counts())
print()
print("El valor medio es ",df_mod['age_rating'].loc[df_mod['age_rating'] > 0].mean(axis=0)," por lo tanto susutituiré las entradas no calificadas por 16")

18    4693
13    2442
15    1654
0      348
17     312
14     146
7        2
Name: age_rating, dtype: int64

El valor medio es  16.044112877067793  por lo tanto susutituiré las entradas no calificadas por 16


df_mod=df_mod.replace(-1, 16)

TFG¶

Indice¶

Introducción ¶

Phase 1 of the CRISP-DM Process Model: Business Understanding ¶

Phase 2 of the CRISP-DM Process Model: Data Understanding ¶

2.1 Gathering data ¶

PENDIENTE: #EXPLICAR EN DETALLE FUNCIONES ¶

2.2 Describing the data ¶

2.3 Exploring data ¶

2.4 Verifying data quality ¶

2.4.1 Unicidad:¶

2.4.2 Completitud:¶

2.4.3 Validez:¶

Phase 3 of the CRISP-DM Process Model: Data Preparation ¶

3.1 Cleaning data¶

3.1.1 Procesado de atributo de edad.¶

3.1.2 Re-escalado de notas de usuarios¶

3.1.3 Limpieza de contenido de reviews con spoilers.¶

3.2 Constructing the clean dataset.¶

Phase 4 of the CRISP-DM Process Model: Modeling ¶

Phase 5 of the CRISP-DM Process Model: Evaluation ¶

Phase 6 of the CRISP-DM Process Model: Deployment ¶

BIBLIOGRAFIA¶

TFG¶

Indice¶

Introducción¶

Phase 1 of the CRISP-DM Process Model: Business Understanding¶

Phase 2 of the CRISP-DM Process Model: Data Understanding¶

2.1 Gathering data¶

PENDIENTE: #EXPLICAR EN DETALLE FUNCIONES ¶

2.2 Describing the data ¶

2.3 Exploring data ¶

2.4 Verifying data quality ¶

2.4.1 Unicidad:¶

2.4.2 Completitud:¶

2.4.3 Validez:¶

Phase 3 of the CRISP-DM Process Model: Data Preparation ¶

3.1 Cleaning data¶

3.1.1 Procesado de atributo de edad.¶

3.1.2 Re-escalado de notas de usuarios¶

3.1.3 Limpieza de contenido de reviews con spoilers.¶

3.2 Constructing the clean dataset.¶

Phase 4 of the CRISP-DM Process Model: Modeling ¶

Phase 5 of the CRISP-DM Process Model: Evaluation ¶

Phase 6 of the CRISP-DM Process Model: Deployment ¶

BIBLIOGRAFIA¶

Introducción ¶

Phase 1 of the CRISP-DM Process Model: Business Understanding ¶

Phase 2 of the CRISP-DM Process Model: Data Understanding ¶

2.1 Gathering data ¶